苹果园区作业机器人障碍物识别研究

Obstacle Recognition in Apple Orchards Robots

苹果采摘的季节性强,工作量大,利用机器人进行苹果采摘符合现代农业的发展趋势。障碍物识别对苹果作业机器人的正常工作至关重要。为此,针对苹果园区垄间地面情况和障碍物特点,采用L*a*b*颜色空间中的b*分量进行图像灰度化处理,再进行图像阈值分割处理,可有效将苹果园区垄间中的障碍物等从复杂背景中分离出来;再利用优化后的SURF算法进行障碍物匹配,可有效地提高障碍物的识别率。实验表明:优化后的SURF算法可以提高障碍物匹配的准确率,并缩短障碍物匹配的时间。

The seasonal picking of apples and the large amount of work required for picking apples using robots are in line with the development trend of modern agriculture. Obstacle recognition is very important for the normal operation of apple working robots. This paper uses the b*component in L*a*b*color space for image gray-scale processing for the ridge floor conditions and obstacle characteristics of apple park. The image threshold segmentation process can effectively separate the obstructions in the ridge of the apple garden from the complex background,and then use the optimized SURF algorithm to perform obstacle matching,which can effectively improve the recognition rate of obstacles. Experiments show that the optimized SURF algorithm can improve the accuracy of obstacle matching and shorten the time of obstacle matching.